In some cases, a system may need to detect whether or not an attachable device is currently present. For example, a server may need to detect whether or not one or more Redundant Array of Independent Disks (RAID) devices are present.
One way of detecting whether or not a device is present is to have the device provide a pre-defined signal sequence to the system. For example, a Serial Advanced Technology Attachment (ATA) will transmit an out-of band signal sequence to indicate its presence in accordance with the Serial ATA Working Group specification entitled “Serial ATA II: Extensions to Serial ATA 1.0” (October, 2002).
FIG. 1 illustrates a portion of a known Serial ATA interface 100 between a disk drive 102 and a system 110. In particular, this portion of the interface 100 is used to provide power to the disk drive 102 via a power supply 120. Note that three conductive paths (i.e., each associated with an interface pin) are provided between the power supply 120 and the disk drive 102 because the amount of current that can be provided through an individual interface pin is limited.
Because a Serial ATA disk drive 102 is “hot-plug” capable (i.e., the disk drive 102 can be attached or removed when power is applied to the system 110) and may include energy storing elements (e.g., capacitors), a large surge of charge-up current could briefly flow from the system 110 to the disk drive 102 when it is attached. To prevent this, one of the ATA interface pins 132 is longer than the other two pins (and will establish contact between the system 110 and the drive 102 before those pins). The conductive path associated with this interface pin 132 includes a resistor 130 having a current limiting resistance of RL.
There is a disadvantage, however, with the use of a signal sequence generated by a device to detect whether or not the device is present. In particular, the system will be unable to detect that a device is present if the device is not operating. For example, a device that is malfunctioning might not be able to generate the appropriate signal sequence, or the interface might not have power applied.
As another approach, a dedicated interface pin could be used to detect whether or not a device is present. For example, a device's interface pin might always be coupled to ground. In this case, a system could detect the presence of the attachment by detecting the presence of ground via a conductive path associated with that interface pin. Requiring a dedicated interface pin, however, could increase the costs associated with an interface. Moreover, such an approach might not even be possible in the case of a pre-defined interface specification.